How to tackle the topic of 'Direction of Flow'

[Nicholas_R_Burns]

Hello Everyone, 

I am reaching out to this community in hopes of learning about any tips or tricks others have used to delineate a 'direction of flow' in some of their utility networks. This could apply to Electrical Distribution, Domestic Water Systems, Chilled Water Systems, Pressurized Steam Systems, etc.

I am new to utility mapping but my goal is to gain some helpful information from some of the more experienced utility experts out there. Our utility assets were recently converted to GIS from a CAD based solution but they still lack a lot of the functionality or supporting attribution that will allow the data to thrive in the GIS environment. At this time, we simply have the line work and points digitized to resemble the previous CAD plans as close as possible. All the data is currently structured in a file geodatabase broken up into datasets for each utility asset. For the issue of flow direction, I used a separate feature class to drop arrows on top of the linework to signify 'direction of flow'. I am aware that this method is tedious and will not allow for future 'upstream/downstream' capabilities when further developing the networks. This is something that I think could eventually add significant value to the data and therefore would like to figure out the best course of action to build this out.

The two main methods I have read about for achieving this are setting 'sinks and sources' or using 'digitized direction of edges'. 

If anyone has experience with these methods and could share some insight, it would be greatly appreciated. Does one method work better than the other? Is one easier to manage down the line? (no pun intended). If I already have the system digitized without consideration of the direction it was drawn in, is there an easy way to set the direction of a line afterwards without having to redraw every line?

Any and all help is greatly appreciated! I look forward to hearing your thoughts.

Thank you,

Nicholas Burns

[MarkBockenhauer]

The flip Iine geoprocessing tool can be used to change the direction of lines that have already been created.  Symbolize the lines with arrows to easily see which way they were digitized.

[MikeMillerGIS]

The Utility Network does not model the concept of flow direction per edge yet.  To calibrate the data, we have a bunch of tools for ArcMap and the Geometric Network.  Since you have the basic network, bring this into a GDB and create a Geometric Network.  You will need to identify a source, but from this source, you can generate direction.  Then use the "Match Selected Lines Direction to Flow Direction" tool.  This will align the digitized direction of the lines to the flow.  

https://solutions.arcgis.com/utilities/water/help/network-editing/tools/overview-network-editing-too...

You could also look at using a Trace Network in ArcGIS Pro.  This uses a Flow Direction network attribute.  This will let you show flow direction arrows based on this attribute.  But since flow direction is not defined from a controller, like a water tower or interconnect, it will be a more manual process then using the GN tools I listed above.

[MikeMillerGIS]

Another thing that can be very helpful, especially when converting from CAD, it is first run the data through a Topology dataset.  You can define rules to ensure data is connected.  It can even help adjust the data to fix where it is not spatially connected.

https://pro.arcgis.com/en/pro-app/latest/help/data/topologies/topology-in-arcgis.htm

[Shawnzeecatz]

We don't use geometric networks and the jury is still out on whether or not we will take the time and effort to do that - I am still very curious as to the benefit of it.

Anyways, for the utilities for which we would like to know/see flow direction, we digitize our lines in the direction of flow. You can 'edit-rt click-flip direction' for any lines that are "flowing" the wrong way. This does take some time if you are just starting out, but once it's done, it's easy to maintain and add new lines from there on out. Then we create a field to calculate the azimuth/bearing direction of that line. Convert those lines to points, symbolize as arrows, and rotate based on that azimuth/bearing direction. Now we have two feature classes, one line and one point. We did this for our stormwater network so we could understand flow through that system - it helped with our MS4 effort. I also automated that process with python so as we update/add lines to the network, I can just run the script to recreate the point layer. Works very well.

It's not a "geometric/utility network" solution, but it works well for us.

If anyone has good reasons why a geometric/utility network would work better, I would honestly love to hear about it.